Original Link: https://www.anandtech.com/show/5785/corsair-performance-series-pro-256gb-review
Corsair Performance Series Pro (256GB) Review
by Kristian Vättö on May 14, 2012 9:10 AM EST- Posted in
- Storage
- SSDs
- Corsair
- Marvell
- Corsair Performance Pro
Introduction
Shortly after our Plextor M3 review went live, I received numerous emails asking us to review Corsair's Performance Series Pro. Your voice was heard and we went and asked Corsair for a review sample, and here we are with the results.
There aren't too many Marvell SSDs on the market so the Performance Pro stole my attention immediately. When testing a SandForce drive, you pretty much know what to expect. Only Intel uses an in-house firmware whereas the rest of the SandForce OEMs are stuck with the firmware that SandForce provides. That limits differentiation a lot. When it comes to Marvell, things are a lot more open and interesting. Firmwares are often proprietary and that's why you never know what to expect.
Personally I'm all for differentiation. The more different SSDs there are, the more competition there is, which is always good from a consumer's standpoint. SandForce is competitive for sure, but most SandForce drives can only compete against each other in price. That's where Marvell and others come in; they offer differentiation in performance and compatibility for instance.
Price competition is not a bad thing but especially since most SandForce OEMs are fabless, it's fairly limited. You buy controllers from SandForce and choose NAND from a few sources; every other SandForce OEM (except Intel) does exactly the same. While every OEM can negotiate their own pricing with SandForce and the selected NAND supplier(s), it's unlikely they will get a significant discount. That's why most SandForce SSDs are priced so similarly. If a certain NAND supplier drops their NAND prices, it always affects more than one SSD OEM.
A Marvell based SSD can still stand out even if it's more expensive since there may not be a comparable drive on the market. Performance is only one aspect; especially garbage collection and power consumption should not go unmentioned as firmware can have a huge impact on them, and hence make drives different.
The Corsair Performance Series Pro
Corsair sampled us with a 256GB unit. Below is a specification table of the Performance Series Pro:
Corsair Performance Series Pro Specifications | ||
Model | 128GB | 256GB |
Raw NAND Capacity | 128GiB | 256GiB |
Usable Capacity | 119.2GiB | 238.5GiB |
Number of NAND Packages | 8 | 8 |
Number of Die per Package | 2 | 4 |
Sequential Read | 500MB/s | 515MB/s |
Sequential Write | 340MB/s | 440MB/s |
4K Random Write | 60K IOPS | 65K IOPS |
Interestingly, Corsair offers only 128GB and 256GB models. I can understand the lack of a 512GB model because of price and low popularity, but 64GB is often one of the most popular models. Corsair does offer 60GB Force Series 3 and Force Series GT drives, and they recently released Accelerator series aimed at caching. There is no specific reason to why Corsair has decided to exclude 64GB from the Performance Series Pro lineup, but it's possible that 64GB was not profitable enough.
64GB SSDs are usually the most expensive in terms of price per GB because the share of NAND in the bill of materials is smaller. In other words, all the other expenses such as controller and manufacturing are the same as in bigger drives. Moreover, 64GB isn't exactly a performance category either. Users who buy such small SSDs are already making a compromise in performance, so they are more likely to grab the cheapest drive instead of paying a bit more for a faster drive.
The Performance Pro does well on paper. Sequential write speeds are very good for a Marvell drive. These days I'm more interested in pricing than the actual specifications, though, mainly because the real world performance difference between most SATA 6Gb/s SSDs is so small that paying more for a slightly faster drive may not be worth it unless your workload is heavily I/O bound. Let's see how Corsair's Performance Series Pro stacks up against other drives in a NewEgg price comparison:
NewEgg Price Comparison (5/14/2012) | ||||
64GB | 128GB | 256GB | 512GB | |
Corsair Performance Series Pro | N/A | $200 | $340 | N/A |
Plextor M3 | $130 | $180 | $340 | $660 |
Crucial m4 | $80 | $120 | $250 | $600 |
Intel 520 Series | $113 | $179 | $331 | $825 |
Samsung 830 Series | $100 | $130 | $310 | $710 |
OCZ Vertex 3 | $165 | $110 | $250 | $650 |
OCZ Vertex 4 | N/A | $150 | $300 | $650 |
The Performance Pro is definitely not the cheapest drive. The 128GB model is actually the most expensive 128GB drive in our comparison and there is $20-40 premium even over the Plextor M3, Intel 520 Series, and Samsung 830 Series, all of which are considered to be high-end drives. The 256GB version is a bit more reasonably priced at $320, although there are still cheaper, competitive drives such as the Samsung 830 Series.
In any case, I would like to point out that SSD prices fluctuate a lot. The price you see today may be different tomorrow. I borrowed the pricing table from our Plextor M3 review and nearly all prices had changed, some even dramatically. If you're buying an SSD, my advice would be to follow the prices for at least a couple of days before pulling the trigger as you may be able to catch a hot sale that way.
Corsair Performance Pro: Packaging and Internals
Corsair Performance Series Pro comes in a compact cardboard box. The box is smaller than most SSD packages I've seen, but it has everything that is necessary: 3.5" bracket, quick installation guide, mounting screws and of course the SSD itself. Unfortunately there is no cloning utility included, which is something you would expect from a high-end drive.
The chassis is made out of brushed metal like most SSDs are nowadays. It feels robust and very well built. Like 2.5" drives in general, the Performance Pro has height of 9.5mm and may hence not be suitable for all ultrabooks.
Heat dissipation is handled by 11 thermal pads: One for the controller, eight for NAND packages and two for DRAMs. The internal layout is a match with Plextor's M3.
From inside, the Performance Pro is identical to the Plextor M3. There are eight quad-die 32GB NAND packages (TH58TEG8D2HBA8C) on the PCB. As with the M3, the NAND is Toggle-Mode MLC NAND manufactured by Toshiba. The DRAM chips are from Nanya and there are two 256MB DDR3-1333 chips on the board. Thermal pads make photographing a bit more difficult since some residue is left on the surface of the chips, so my apologies if you can't see the markings clearly.
On the other side of the PCB, we find Marvell's 88SS9174-BLD2 controller, again exactly the same as used in Plextor M3. Is it just a coincidence that these two SSDs are so similar or is there a reason behind it? It's time to find out!
Plextor M3 Pro and M3 on the top and Corsair Performance Series Pro underneath the two
Neither Plextor or Corsair manufactures the drives they sell. Both companies buy their drives from Lite-On and if you take a look at Lite-On's M3S SSD, it seems a lot like Corsair's Performance Series Pro and Plextor's SSDs, doesn't it? That's why the hardware of the Performance Pro and M3 is the same—they come from the same plant. However, it should be kept in mind that hardware is only one part of the equation.
As you will see in our tests, the Performance Pro and M3 do not perform the same. This is due to firmware differences. Corsair uses a stock firmware provided by Marvell while Plextor has a whole R&D team dedicated to their SSD firmware development. In other words, Plextor uses an in-house firmware which is also the reason for differences in performance. Whether Plextor's firmware is a tweaked stock firmware or totally built from a scratch I don't know, but either way it's not the same as what Corsair uses.
The Test
CPU |
Intel Core i5-2500K running at 3.3GHz (Turbo and EIST enabled) |
Motherboard |
AsRock Z68 Pro3 |
Chipset |
Intel Z68 |
Chipset Drivers |
Intel 9.1.1.1015 + Intel RST 10.2 |
Memory | G.Skill RipjawsX DDR3-1600 2 x 4GB (9-9-9-24) |
Video Card |
XFX AMD Radeon HD 6850 XXX (800MHz core clock; 4.2GHz GDDR5 effective) |
Video Drivers | AMD Catalyst 10.1 |
Desktop Resolution | 1920 x 1080 |
OS | Windows 7 x64 |
Random Read/Write Speed
The four corners of SSD performance are as follows: random read, random write, sequential read and sequential write speed. Random accesses are generally small in size, while sequential accesses tend to be larger and thus we have the four Iometer tests we use in all of our reviews. Our first test writes 4KB in a completely random pattern over an 8GB space of the drive to simulate the sort of random access that you'd see on an OS drive (even this is more stressful than a normal desktop user would see).
We perform three concurrent IOs and run the test for 3 minutes. The results reported are in average MB/s over the entire time. We use both standard pseudo randomly generated data for each write as well as fully random data to show you both the maximum and minimum performance offered by SandForce based drives in these tests. The average performance of SF drives will likely be somewhere in between the two values for each drive you see in the graphs. For an understanding of why this matters, read our original SandForce article.
Performance Pro's random read performance is slightly faster than Plextor's M3's but there is essentially no difference. In general the random read performance is excellent, as expected.
Random write performance is not top-notch but still ~6% faster than what Plextor's M3 provides.
Increasing the queue depth shows an increase in random write performance but this time the M3 is a bit faster. We are only looking at a difference of ~4% though, and it should be noted that most workloads don't use queue depths higher than 5.
Sequential Read/Write Speed
To measure sequential performance we ran a one minute long 128KB sequential test over the entire span of the drive at a queue depth of 1. The results reported are in average MB/s over the entire test length.
Again, identical performance with Plextor M3 and this time Crucial's m4 joins the chorus as well.
Sequential write introduces some surprises. Plextor M3 is 34MB/s (10.7%) faster here while in other tests the two were much closer to each other. However, one shouldn't pay too much attention to just one test and our Storage Suites should give a much better look at real world performance.
AS-SSD Incompressible Sequential Performance
The AS-SSD sequential benchmark uses incompressible data for all of its transfers. The result is a pretty big reduction in sequential write speed on SandForce based controllers, while other drives continue to work at roughly the same speed as with compressible data.
Incompressible sequential read testing is fairly boring as nearly all SSD capable of SATA 6Gb/s perform the same, and the Performance Pro is no exception.
Incompressible sequential write speed is another story, though. Here the Performance Pro is dominating our charts. It's actually surprising how big the difference is. We are looking at over 60MB/s (17.2%) difference between Plextor M3 and the Performance Pro.
AnandTech Storage Bench 2011
Last year we introduced our AnandTech Storage Bench, a suite of benchmarks that took traces of real OS/application usage and played them back in a repeatable manner. Anand assembled the traces out of frustration with the majority of what we have today in terms of SSD benchmarks.
Although the AnandTech Storage Bench tests did a good job of characterizing SSD performance, they weren't stressful enough. All of the tests performed less than 10GB of reads/writes and typically involved only 4GB of writes specifically. That's not even enough exceed the spare area on most SSDs. Most canned SSD benchmarks don't even come close to writing a single gigabyte of data, but that doesn't mean that simply writing 4GB is acceptable.
Originally we kept the benchmarks short enough that they wouldn't be a burden to run (~30 minutes) but long enough that they were representative of what a power user might do with their system. Later, however, we created what we refer to as the Mother of All SSD Benchmarks (MOASB). Rather than only writing 4GB of data to the drive, this benchmark writes 106.32GB. This represents the load you'd put on a drive after nearly two weeks of constant usage. And it takes a long time to run.
1) The MOASB, officially called AnandTech Storage Bench 2011—Heavy Workload, mainly focuses on the times when your I/O activity is the highest. There is a lot of downloading and application installing that happens during the course of this test. Our thinking was that it's during application installs, file copies, downloading, and multitasking with all of this that you can really notice performance differences between drives.
2) We tried to cover as many bases as possible with the software incorporated into this test. There's a lot of photo editing in Photoshop, HTML editing in Dreamweaver, web browsing, game playing/level loading (Starcraft II and WoW are both a part of the test), as well as general use stuff (application installing, virus scanning). We included a large amount of email downloading, document creation, and editing as well. To top it all off we even use Visual Studio 2008 to build Chromium during the test.
The test has 2,168,893 read operations and 1,783,447 write operations. The IO breakdown is as follows:
AnandTech Storage Bench 2011—Heavy Workload IO Breakdown | ||||
IO Size | % of Total | |||
4KB | 28% | |||
16KB | 10% | |||
32KB | 10% | |||
64KB | 4% |
Only 42% of all operations are sequential; the rest ranges from pseudo to fully random (with most falling in the pseudo-random category). Average queue depth is 4.625 IOs, with 59% of operations taking place in an IO queue of 1.
Many of you have asked for a better way to really characterize performance. Simply looking at IOPS doesn't really say much. As a result we're going to be presenting Storage Bench 2011 data in a slightly different way. We'll have performance represented as Average MB/s, with higher numbers being better. At the same time we'll be reporting how long the SSD was busy while running this test. These disk busy graphs will show you exactly how much time was shaved off by using a faster drive vs. a slower one during the course of this test. Finally, we will also break out performance into reads, writes, and combined. The reason we do this is to help balance out the fact that this test is unusually write intensive, which can often hide the benefits of a drive with good read performance.
There's also a new light workload for 2011. This is a far more reasonable, typical every day use case benchmark. It has lots of web browsing, photo editing (but with a greater focus on photo consumption), video playback, as well as some application installs and gaming. This test isn't nearly as write intensive as the MOASB but it's still multiple times more write intensive than what we were running last year.
We don't believe that these two benchmarks alone are enough to characterize the performance of a drive, but hopefully along with the rest of our tests they will help provide a better idea. The testbed for Storage Bench 2011 has changed as well. We're now using a Sandy Bridge platform with full 6Gbps support for these tests.
AnandTech Storage Bench 2011—Heavy Workload
We'll start out by looking at average data rate throughout our new heavy workload test:
Samsung SSD 830 is still unbeatable but the Performance Pro turns out to be the fastest Marvell based SSD. It's only 3.7% faster than Plextor M3 though so the difference is not significant.
While Performance Pro was slower in sequential write performance when tested with compressible data, its outstanding incompressible sequential write speed makes it a bit faster than Plextor's M3. This is why it's always important not to look at just one benchmark when estimating the overall performance.
The next three charts just represent the same data, but in a different manner. Instead of looking at average data rate, we're looking at how long the disk was busy for during this entire test. Note that disk busy time excludes any and all idles; this is just how long the SSD was busy doing something:
AnandTech Storage Bench 2011, Light Workload
Our new light workload actually has more write operations than read operations. The split is as follows: 372,630 reads and 459,709 writes. The relatively close read/write ratio does better mimic a typical light workload (although even lighter workloads would be far more read centric). The I/O breakdown is similar to the heavy workload at small IOs, however you'll notice that there are far fewer large IO transfers:
AnandTech Storage Bench 2011—Light Workload IO Breakdown | ||||
IO Size | % of Total | |||
4KB | 27% | |||
16KB | 8% | |||
32KB | 6% | |||
64KB | 5% |
As our Light Storage Suite is more small IO centric than our Heavy suite, the difference between the Performance Pro and Plextor M3 is only 2MB/s now, which is essentially nothing.
Performance Over Time and TRIM
Garbage collection and TRIM are often things that separate very similar drives if they have different firmware. I ran a few more tests than we usually run in order to understand the drive's behavior better and hopefully illustrate some questions that our readers have brought up. Let's start off by running HD Tach on a clean drive so we have a reference for comparison:
To begin our torture session, I filled the drive with compressible data and proceeded to hammer it with compressible 4KB random writes (QD32, 100% LBA space) for 20 minutes:
We are looking at 171.6MB/s on average, although the performance does drop to as low as ~60MB/s on the earliest LBAs. For comparison, Plextor's M3 scored an average of 163MB/s so the two are very close to each other once again.
As you can see, 20 minutes was not enough to put the Performance Pro on its knees, thus I secure erased the drive, filled it again and now ran our torture test for 60 minutes:
That's more like it. Average write speed drops to 57.4MB/s and at the lowest the Performance Pro is only able to write at ~35MB/s.
The next graph is HD Tach after 60 minutes of torture, followed by 40 minutes of idle time. Note that I did not run HD Tach in between, the drive was left idle right after torturing.
And the performance is almost the same as without idle time. There is around 4MB/s difference in average write speed, which is essentially nothing.
However, there is a rescue: sequential writes. The graph below is also after 40 minutes of idle time but now I ran HD Tach right after the 60 minute torture and then let the drive idle for 40 minutes:
Performance is much better now. The purpose of this test is to illustrate how sequential writes help to restore performance. It's actually quite logical. If you have a drive full of random data and another write comes in, there is a good chance that read-modify-write will be triggered (i.e. the block must first be read to cache and then rewritten with the new page of data). As Anand explained in our SSD Anthology, read-modify-write operations cause a decline on write speeds from a user's perspective—the actual drive is still writing at the same speed but now it has to read the block first, modify select pages and then rewrite the whole block.
On the other hand, a drive that's full of random data which was overwritten by sequential data effectively looks like a post-TRIM drive. The controller can simply throw away all dirty blocks and start writing from scratch. There is still some tracking that the controller needs to do and read-modify-write may be triggered, hence there's lower performance than at clean state, but there is a lot less compared to a drive full of random data.
I decided to run HD Tach one more time after the drive had been idling for two hours:
And performance is almost back to new. Note that this is the third time I'm running HD Tach after the drive was secure erased and tortured, so that's even more sequential writes.
Finally, I secure erased the drive, tortured for 60 minutes and then TRIM'd to see that TRIM is effective—and it is.
To analyze these results quickly, there is once again absolutely no problem if you're running an OS with TRIM support. I would also say that the Performance Pro is fine for an OS without official TRIM support (e.g. OS X). As the graphs show, performance can get pretty bad if the drive is put into a tortured state, but this is the case with all SSDs we have seen so far. However, idle time and sequential writes come to the rescue.
Also, TRIM can nowadays be enabled in OS X. Even if you are not comfortable with TRIM being enabled 24/7 since it's not officially supported by Apple for third party drives, you can always enable TRIM, secure erase free space in Disk Utility, and then disable TRIM. That will TRIM the empty space in your SSD and hence restore performance.
Furthermore, our tests are extreme and their purpose is to show how the drive behaves when it's backed into a corner. You would need a very heavy and specific workload to put the drive in this state, so especially for a normal user there is nothing to worry about even if you're running OS X.
Power Consumption
Generally speaking Marvell SSDs do quite well in power consumption tests, so I was interested to see if Corsair is able to beat Plextor's M3, given that Plextor uses a custom firmware whereas Corsair does not.
Idle power consumption is identical to Plextor's M3 and they are both the most power efficient drives at idle of what we have tested.
The Performance Pro is also the most power efficient non-SF drive in sequential writes.
All in all, the Performance Pro is extremely power efficient and is overall definitely one of the most efficient drives we have tested. It would certainly make a good fit for a laptop drive.
Final Words
Corsair's Performance Series Pro is very fast, and well behaved. Its power consumption is also nice and low, making it a good fit for notebook users. In the end, it all boils down to pricing. SSD pricing is highly volatile - at the time of publication the Performance Series Pro, at least at Newegg, came in a bit too high.
Given the solid track record of Samsung's SSD 830 and its good performance, we really need to see pricing more in-line with it or Crucial's m4 for the Performance Series Pro to make sense. If you come across one of these drives on a good sale however, don't hesitate to pull the trigger.
Longer term there are always concerns about Marvell's ability to deliver timely firmware updates to its customers who lack their own internal firmware teams. It's because of this that pricing is even more important to get right.
NewEgg Price Comparison (5/14/2012) | ||||
64GB | 128GB | 256GB | 512GB | |
Corsair Performance Series Pro | N/A | $200 | $340 | N/A |
Plextor M3 | $130 | $180 | $340 | $660 |
Crucial m4 | $80 | $120 | $250 | $600 |
Intel 520 Series | $113 | $179 | $331 | $825 |
Samsung 830 Series | $100 | $130 | $310 | $710 |
OCZ Vertex 3 | $165 | $110 | $250 | $650 |
OCZ Vertex 4 | N/A | $150 | $300 | $650 |